Abstract
This paper studies how changes in land prices affected spatial population movement in Tokyo from 1983 to 2006. We first estimate land prices in Tokyo by universal cokriging, which capitalizes on the spatial and temporal correlations in land prices. We next build a structural equation model (SEM) for migration among the municipalities in Tokyo using economic variables, including land prices. The Kullback-Leibler (KL) information for all combinations of two municipalities in Tokyo is then computed from the estimated SEM. Finally, we classify the municipalities in Tokyo based on the KL information. The resultant four groups are characterized from an economic viewpoint. More applications are needed to test the reliability of our approach.
